Date: Wed, 1 Mar 2000 16:33:31 -0800 (PST)
From: Linus Torvalds <torvalds@transmeta.com>
On Thu, 2 Mar 2000, Andrea Arcangeli wrote:
> It's softnet. I'll do a few traces now and I'll let you know what's going
> on. Thanks!
Softnet is certainly there too. But I thought that Lawrence was trying the
filesystem-only portion of the simulated dbench.
Lawrence? Are you using the full dbench with sockets and networking, or
are you just replaying the IO part of it?
I think he's just doing the disk portion of dbench.
Softnet could be to blame in that there may be a salient
bad side-effect to how "classical BH" is implemented in terms
of the new software irqs.
Is everyone seeing this problem using SCSI? That would be a good
reference point, because SCSI command done processing is still
implemented as a "classical BH".
One would also wish to take a look at how the various task queues
behave in the new code as well.
Later,
David S. Miller
davem@redhat.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Mar 07 2000 - 21:00:11 EST